In the midst of what some call a "crisis" in the ability to replicate results of laboratory studies, a European consortium has conducted a preclinical randomized controlled trial (pRCT), which mimics the design and rigorous standards of a clinical trial but is done in animals, and provides one more view of the drug's efficacy before it is used in the clinical trial setting.
The study evaluated a potential stroke therapy that had proved effective in one mouse model but not another, yet is already in phase 2 clinical trials.
"I think most researchers in the field--including me--had the feeling that this clinical study was somewhat prematurely initiated," said the study's senior author Arthur Liesz, from the University Medical Center Munich.
Results of Liesz and his team suggest that the success of the stroke therapy may depend on the type and severity of stroke, knowledge that could have improved clinical trial design for this therapy.
More broadly, the pRCT may offer a potentially powerful tool for bridging the gap between the lab and clinic and could help select the most promising drug candidates before launching costly and potentially harmful clinical trials.
"Our study unequivocally proved that preclinical randomized multicenter trials are feasible, which has been doubted over the last decade," said the study's first author, Gemma Llovera, from the University Medical Center Munich, Germany.
"We believe that pRCTs might become a new step within the pipeline from basic research [to] preclinical drug testing to clinical trials," said Liesz.
The findings appear in the 5 August issue of Science Translational Medicine.
Biomedical research is widely known to be plagued by the inability to replicate results of preclinical studies, also known as the "reproducibility crisis." Many treatments that show promise in the lab fail in clinical trials, often due to limited sample sizes, flaws in statistical analysis, inadequate reporting of data, or lack of blinding or randomization.
"[The research community is] confronted with the situation that none--out of hundreds of preclinically successfully tested drug candidates--was shown to be effective in a clinical trial," said Liesz.
A preclinical trial has been proposed as a way to address the replication crisis. It is modeled after a clinical randomized controlled trial (RCT), the gold standard in clinical drug development. International research consortia have already been established with the goal of performing pRCTs, including the European Union-funded Multicentre Preclinical Animal Research Team (Multi-PART) for stroke research and the U.S. National Institutes of Health-funded Consortium for preclinicAl assESsment of cARdioprotective therapies (CAESAR) for cardiac research.
The European consortium led by Liesz investigated a candidate stroke therapy called CD49d-specific antibodies. Four rodent studies previously showed that these antibodies protected the brain from inflammation after ischemic stroke, but were contradicted by a fifth study showing no benefit. Nevertheless, a phase 2 clinical trial of the antibody therapy, which has been approved to treat multiple sclerosis, was launched in stroke patients.
To determine whether a pRCT could have better guided the decision to advance the anti-CD49d antibody to clinical trials, six independent research centers in Europe collaborated to conduct a multicenter, centrally coordinated, blinded pRCT using two different mouse models of stroke. Each lab performed experiments using similar rigorous standard procedures and many of the same materials, down to the surgical equipment and mice of identical age and gender from the same commercial breeder. Results from a total of 315 mice were pooled from all centers.
"[We were] learning and adopting key measures from current clinical trial design [to bring] our pRCT closer to the design and rigor of clinical trials," said Liesz.
In a mouse model of confined stroke, the anti-CD49d antibody therapy stemmed inflammation and reduced brain damage. In another model of severe stroke, which caused more extensive brain damage, the antibody showed no protective effect. The results suggest that the success of the treatment may depend on the type and severity of the stroke, which could inform the design of future studies.
The trial, from initiation to unblinding, was performed over the course of a year and four months and cost a total of $180,000 or €165,000, including more than €30,000 direct drug costs.
"Compared to the enormous financial and scientific loss caused by a negative clinical trial, preventing a single unnecessary or even harmful clinical trial by conducting a pRCT to confirm single center preclinical results will outweigh the costs and time of a pRCT manifold," said Liesz.
The researchers plan to compare their findings with the outcome of the recently completed phase 2 clinical trial, expected to be released later this year.
Altogether, the study demonstrates the feasibility of conducting a pRCT and provides a road map for future preclinical trials. It also uncovers unexpected variability between the two commonly used stroke models and weak sensitivity of certain behavioral tests. Future studies could pool results from even more models and validate some of the methods before performing a multicenter trial, the researchers say.
"The future pRCTs should be potentially conducted in close collaboration with pharmaceutical industry in order to improve translation of study design and outcome parameters to clinical trials," said Liesz. "If supported by funding agencies and the pharmaceutical industry, pRCTs [may] become a mandatory step before testing a drug candidate in a clinical trial."
###
Journal
Science Translational Medicine